This is a parallel application from researchers at the Population Research Centers of the University of Wisconsin and University of Maryland. The overall goal of this application is to examine patterns, transitions, and determinants of health in a comparative framework within countries of the Latin American and Caribbean region. The study seeks to contrast patterns within the region and also with those found in the U.S., in particular with populations of Mexican origin, taking advantage of recently gathered, high- quality, comparable, and as of yet under-analyzed survey data on elderly people. The goal is to exploit the available cross-country data to identify conditions that promote (inhibit) the emergence of relations between health status and disability and well-specified determinants, and to assess how recurrent and generalizable are relations found in any one country. The specific aims of the project are: to estimate profiles and determinants of health status and disability of elderly populations in Mexico, and in seven capital cities of Latin America (Argentina, Barbados, Brazil, Cuba, Chile, Mexico, and Uruguay); to identify determinants of health and disability status using cross-sectional data to contrast health and disability profiles of populations in Mexico with that of Mexicans and Mexican-origin populations living in the U.S.; to estimate correlates of health status changes, disability transitions and mortality in Mexico and the U.S. with particular emphasis on the role of early childhood conditions, experiences with illness, and on the influence of income and wealth. Because of their increasing importance in the region, the project focuses more narrowly on diabetes and obesity, and proposes to estimate models for incidence and duration of diabetes as well as health-related costs associated with these conditions. In pursuing these goals the study adds to and fine-tunes existing methods and procedures to help address two issues: health selection among immigrants to the U.S., and estimation of couple-models to assess the effects of select determinants on health and mortality net of shared attributes. [unreadable] [unreadable] [unreadable] [unreadable]